In Germany there's a connection between FiT rates and the price of PV systems. Market forces have conspired in such a way that the IRR for residential installs stays relatively steady at around 8%. Given a reasonable set of assumptions we can guess what system prices will do as the FiT falls. In my experience you get surprisingly accurate projections for how average system costs will change. Last year around this time I sketched out how I saw Germany's installation costs trending during the next year. I figured prices would trend to around 2500 Euro/kW for 2011. According to the BSW, Q2 average prices are 2422 Euro/kW so my guess work was pretty good.

Current average price: ~2422 Euro/kW

Estimated LCOE: ~19 cents/kWh

Current best FiT: 28.74 cents/kWh

Last year's sketch was done before Germany set in motion plans to accelerate the FiT reduction in the second half of 2011. Here's the projection for the second half.

Steepest case scenario: If the FiT drops to 24.43 cents/kWh in the second half you'd expect system prices to get down to 2000 to 2100 Euro/kW. Jumping forward to 2012 the worst case schedules imply system costs will come down to a range of 1800 to 1900 Euro/kW. I don't actually expect a steepest case scenario. These numbers are for benchmarking only at this point.

One thing to look out for is how retail electricity prices should start buoying system prices next year in an appreciable way. Consumers are increasingly running into this sort of choice.

1. Selling the kWh for a profit of 7 cents

2. Using the kWh for a savings of 10 cents

This spread creates an extra profit potential for those willing to manage their energy use. How big is this incentive? Some behavioral changes should come naturally. For example, in contrast to what we've normally heard about running appliances in off-hours, the German FiT structure will increasingly incentivize PV owners to run appliances during the day where possible. For the sake of argument, let's say there's 100 Euros of annual savings that can be captured by consumers who purchase smarter appliances and make minor adjustments to when they do the laundry or run the dishwasher. This isn't a whole lot of money but my thinking is, well hell, recycling is more of a pain in the ass and saves me less. This seems like it has potential so I've been trying to dig into the question.

How do you more accurately quantify the possibility here? To understand the problem it helps to step back a few years. Before the self-consumption kicker was incorporated into Germany's FiT schedule it was more straight-forward to model the profitability of a PV system. The expected annual production, interest rate, FiT rate, discount rate, maintenance and insurance costs were the primary variables that determined profitability. The self-consumption kicker creates a more difficult modeling procedure. You can't simply look at annual production - you have to look at your hourly production and determine what percentage of that production will be self-consumed. To try to solve this problem you can:

1. Compare historic weather patterns against historic load patterns

2. Run a synthetic weather generator against a synthetic load generator

3. Use a thumb rule guess and assume that 30% of production will be self-consumed

My estimation procedure has always used option three but I think option two validated against option one is the optimum procedure. Although one might well exist, I don't know of a software package that models self-consumption directly. That said, there are definitely individual packages that synthetically model weather and/or load individually - i.e. Tools already exist to model this problem.

What would you want to find with your modelling? As a first order of business you want to find out what your natural self-consumption is. Natural self-consumption is the amount of consumption that "naturally" lines up with production. No fancy appliances required, no behaviors are changed. Once you have this baseline established you can start modeling managed self-consumption.

How could you manage consumption? One possibility is heat pump water heaters. This appliance has two attributes which allow it to help manage self-consumption. 1. HPWHs use electricity 2. HPWHs inherently store a product (hot water) which means you don't have to buy an extra storage device. If, as described above, you can model when you think your PV system will be producing against when you think you'll need the hot water you can estimate your technical potential for shifting. How much of this technical potential can be captured will be determine by running an IRR of the cost of shifting vs. the savings of shifting. Honestly, I don't know the answer to this question - not yet at least.

All this Goldbergian musing is established on the idea that we'll see a sizeable spread between the price one gets for taking the FiT as opposed to the savings captured by using the electricity instead. This spread will exist even after the self-consumption kicker is discontinued and, more importantly, it should increase as the FiT and retail electricity rates continue to diverge away from each other in 2012 and beyond. Will it be big enough to drive hot water arbitrage? I think there's a good chance.

One other random thought... PV electricity allows one to escape grid fees but it doesn't make the grid fees go away. As Germany adds more and more PV the grid fees will have to be divided over fewer kWhs sold. How much will this make electricity rates go up? Divide grid fees by total kWhs sold today compared to total kWhs expected to be sold at higher PV penetrations and you've got a good estimate.